As Cronbach's Alpha is usually used for test the reliability of measurements given by questionnaires, however, I wonder that whether I use this test for features extracted from behavioral data.
I have a system for user to commit artefacts of their project iteratively (freely resubmission until all requires are satisfied) , as I would like measure their productivity by number of submissions, number of generated errors, number of solved errors... totally about ten statistical features can extracted from the log data. I would like to measure the productivity/effort of user by this ten features. In the meantime, I did the test and got the Alpha score 0.81. All features are ok for my further hypothesis, however, should I add a results of Cronbach's Alpha test into the paper?
Given that you have already computed it and the results fit the purpose of the paper, you might include it.
From what I know and recall the Cronbach alpha is often / usually used with ordinal and/ or nominal data. You could report these results but also use factor analysis or principal components analysis as additional measurements, especially if most of these variables are ratio or interval based.
- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts